|LL Parsing email@example.com (R. Anthony Vestal) (2000-10-19)|
|Re: LL Parsing LLkParsing@aol.com (2000-10-22)|
|Re: LL Parsing firstname.lastname@example.org (2000-10-22)|
|Re: LL Parsing email@example.com (Randall Hyde) (2000-10-23)|
|Date:||22 Oct 2000 01:22:02 -0400|
|Organization:||Deja.com - Before you buy.|
> Are LL(k) parsers still used in medium to large applications?
I think that "starting to be" is the proper question rather than
"still", assuming that k > 1. The problem with LL(1) is that its
expressive power is very limited. In the past few years, several
parser generators have appeared that can handle limited forms of
LL(k). The advantages of LL parsing are
- It mirrors recursive programming, which is easier to understand than
the state machine oriented LR methods.
- Better error detection.
- Inherited attributes are handled naturally.
LR parsing advantages:
- Only needs a single lookahead level, LALR(1) is powerful enough.
- Handles synthesized attributes easily.
- Attributes can be kept on the parse stack.
- Mature tools like YACC are widely available.
Return to the
Search the comp.compilers archives again.